Promotion of Plasmalogen Biosynthesis Reverse Lipid Changes in a Barth Syndrome Cell Model T ⁎ José Carlos Bozelli Jr.A, Daniel Lub, G

BBA - Molecular and Cell Biology of Lipids 1865 (2020) 158677

Contents lists available at ScienceDirect

BBA - Molecular and Cell Biology of Lipids

journal homepage: www.elsevier.com/locate/bbalip

Promotion of plasmalogen biosynthesis reverse lipid changes in a Barth Syndrome cell model T ⁎ José Carlos Bozelli Jr.a, Daniel Lub, G. Ekin Atilla-Gokcumenb, Richard M. Epanda, a Department of Biochemistry and Biomedical Sciences, McMaster University, Health Sciences Centre, Hamilton, Ontario L8S 4K1, Canada b Department of Chemistry, University of Buﬀalo, The State University of New York, Buﬀalo, NY 14260, USA

ARTICLE INFO ABSTRACT

Keywords: In Barth syndrome (BTHS) mutations in tafazzin leads to changes in both the quantities and the molecular Barth syndrome species of cardiolipin (CL), which are the hallmarks of BTHS. Contrary to the well-established alterations in CL Plasmalogen associated with BTHS; recently a marked decrease in the plasmalogen levels in Barth specimens has been Cardiolipin identified. To restore the plasmalogen levels, the present study reports the effect of promotion of plasmalogen Phospholipidomics biosynthesis on the lipidome of lymphoblasts derived from Barth patients as well as on cell viability, mi- Mass spectrometry tochondria biogenesis, and mitochondrial membrane potential. High resolution 31P NMR phospholipidomic Mitochondrial properties analysis showed an increase in the levels of plasmenylethanolamine (the major plasmalogen in lymphoblasts), which reached values comparable to the control and a compensatory decrease in the levels of its diacyl-PE counterpart. Importantly, 31P NMR showed a significant increase in the levels of CL, while not altering the levels of monolysocardiolipin. Mass spectrometry measurements showed that the promotion of plasmalogen biosynthesis did not change the molecular species profile of targeted phospholipids. In addition, promotion of plasmalogen biosynthesis did not impact on cellular viability, although it significantly decrease mitochondria copy number and restored mitochondrial membrane potential. Overall, the results showed the efficacy of the promotion of plasmalogen biosynthesis on increasing the CL levels in a BTHS cell model and highlight the potential beneficial effect of a diet supplemented with plasmalogen precursors to BTHS patients.

1. Introduction peripheral membrane, instead of parallel cristae that often extend across the entire organelle observed in controls [9]. In addition, CL also Barth Syndrome (BTHS) is an X-linked recessive disease caused by plays a role in keeping the integrity of the inner mitochondrial mem- mutations in the gene that encodes tafazzin [1,2]. Tafazzin is a lyso- brane (IMM) and alterations that compromise it would result in dis- phospholipid-phospholipid transacylase involved in the last step of sipation of the proton gradient that drives ADP phosphorylation. In- cardiolipin (CL) biosynthesis [3,4]. In BTHS loss-of-function of tafazzin deed, BTHS mitochondria exhibit increased proton leakage, decreased leads to a lower content of CL as well as more heterogeneous molecular membrane potential, and a decreased respiratory coupling index species of this lipid and accumulation of monolysocardiolipin (MLCL), [9–12]. Moreover, in Barth specimens, changes in CL lead to destabi- which are considered the hallmark of BTHS [5–7]. In eukaryotes, CL is a lization of mitochondrial super complexes (SC) [9,13,14]. Likewise, mitochondria-speciﬁc lipid, which has been reported to play crucial assembly of mitochondrial SC induces CL remodeling and, as a con- roles in the structure and function of mitochondria [8]. Hence, it is not sequence, its stabilization [15]. Mitochondria are key players in cellular unexpected that a major focus of research in this disease has been to redox metabolism and disruption of mitochondrial SC is known to result understand the molecular events resulting from changes in both the in increased production of reactive oxygen species (ROS) [16,17]. In- quantities and the molecular species of CL on mitochondrial structure deed, in several BTHS specimens elevated ROS production has been and function. reported [9,12,18,19]. Cells from BTHS patients usually display in- In BTHS mitochondrial morphology is aberrant, presenting swollen creased mitochondria biogenesis as indicated by the rise in the AMP/ and enlarged mitochondria with malformed cristae structures, which ATP ratio and mitochondria copy number when compared to controls, have lower surface area and appear round and disconnected from the which is proposed to occur for cellular compensation due to the

⁎ Corresponding author at: Department of Biochemistry and Biomedical Sciences, McMaster University, 1280 Main Street West, Hamilton, Ontario L8S 4K1, Canada. E-mail address: [email protected] (R.M. Epand). https://doi.org/10.1016/j.bbalip.2020.158677 Received 20 November 2019; Received in revised form 6 February 2020; Accepted 27 February 2020 Available online 29 February 2020 1388-1981/ © 2020 Elsevier B.V. All rights reserved. J.C. Bozelli, et al. BBA - Molecular and Cell Biology of Lipids 1865 (2020) 158677 impaired mitochondria [9,11]. the efficacy of the promotion of plasmalogen biosynthesis on increasing While the alterations in CL associated with BTHS are well-estab- the CL levels in a BTHS cell model and suggest that a diet supplemented lished, currently there has been an increased awareness of more with plasmalogen precursors could be beneficial for BTHS patients. widespread changes in the lipid composition of cells depleted of tafazzin function. Of note, it is the recent identification of a decrease in the plasmalogen levels in organs obtained from tafazzin knockdown 2. Materials and methods mice as well as blood cells derived from BTHS patients, which clearly indicate that plasmalogen loss is associated with BTHS [20,21]. 2.1. Materials Plasmalogens are members of a subclass of glycerophospholipids, which differ from their diacylglycerophospholipid counterpart by Sodium dodecyl sulfate (SDS) was from Bioshop (Canada). 2-etha- having a fatty alcohol linked via a vinyl-ether bond at the sn-1 position nesulfonic acid (MES), ethanolamine, CsOH, butylated hydroxytoluene of the glycerol moiety rather than an esterified fatty acid. Plasmalogens (BHT), ethylenedinitrilo-tetraacetic acid (EDTA), bovine serum albumin fl comprise up to 20% of the total phospholipid mass in humans and are (BSA), carbonyl cyanide p-(tri uoromethoxy)phenyl-hydrazone constituents of the plasma membrane as well as intracellular mem- (FCCP), methanol (HPLC grade), and chloroform (HPLC grade) were branes (nucleus, endoplasmic reticulum, post-Golgi network, mi- from Sigma-Aldrich (St. Louis, MO). D2O (99.9 atom %D) was from tochondria) [22]. Plasmenylethanolamine (PE-Pls) is the predominant Cambridge Isotope Laboratories (Tewksbury, MA). Roswell Park plasmalogen found in most of human cells, except in cardiac and ske- Memorial Institute (RPMI) 1640 Medium was from Gibco letal muscles and mature spermatozoa where high proportions of both (Gaithersburg, MD). Fetal bovine serum (FBS) was from Wisent PE-Pls and plasmenylcholine are found [22]. Although the physiolo- (Canada). 1-O-hexadecyl-sn-glycerol (HG) was from Avanti Polar Lipids gical role of plasmalogen remains elusive, several studies had shed light (Alabaster, AL). All reagents were used as received. on some of their physical, chemical, and biological functions. For in- stance, it has been reported that the vinyl-ether bond at the sn-1 posi- 2.2. Cell culture and plasmalogen precursor administration tion of the glycerol moiety brings the aliphatic chains closer together in comparison to their diacyl counterparts [23,24]. This leads to more Lymphoblast cell lines were established by Epstein–Barr virus fl packed membranes with decreased uidity and increased order, while transformation of the leukocytes isolated from the whole blood of two more easily promoting the formation of non-bilayer phases. In addition, BTHS patients and their gender- and age-matched control subjects using the vinyl-ether bond is oxidation-labile rendering plasmalogens the Ficoll–Hypaque gradients as previously described [11]. BTHS and ability to scavenger ROS, which led to the proposition they could act as control lymphoblasts were cultured in RPMI 1640 medium in the pre- a protection against the oxidation of polyunsaturated fatty acids sence of 10% FBS, penicillin (50 IU/mL), and streptomycin (50 μg/mL) (PUFAs) and other oxidation prone lipids [25–27]. Moreover, plasma- at 37 °C in a humidified 5% CO2 atmosphere. Cells were seeded at a logens are believed to act as reservoirs of second messengers as they density of ca. 2×105 cells/mL and passaged every 2–3 d. Adminis- are, usually, enriched with PUFAs at the sn-2 position of the glycerol tration of plasmalogen precursors was conducted by adding 30 μMof moiety, which are biologically active lipid mediators released upon HG and 5 μM of ethanolamine (this component was added to avoid it plasmalogen hydrolyses by PLA2 [22]. becoming rate limiting for the synthesis of PE-Pls) 20 h before har- A decrease in the steady-state levels of plasmalogens is associated vesting the cells [30]. Other two concentrations (2.5 fold above and with an impaired biosynthesis or an enhanced degradation. Currently below the one mentioned) of HG and ethanolamine were tested, but the mechanism of plasmalogen depletion in BTHS at the molecular level data is only presented for the above concentrations as they yielded the is not known. However, in several diseases characterized by depletion better recovery of PE-Pls and CL. HG was dissolved in ethanol and the in plasmalogens, promotion of their synthesis has been shown as a final ethanol concentration in the cell suspension was 0.25% (v/v). successful strategy to alleviate/treat those conditions [22]. Further- Administration of vehicle to control and Barth lymphoblasts did not more, disruption of plasmalogen biosynthesis impaired mitochondrial change significantly the phospholipids as measured by 31P NMR in ei- dynamics, which could be rescued upon promotion of plasmalogen ther case (data not shown). Cells were harvested by centrifugation at biosynthesis [28]. While restoring plasmalogen levels is relatively 1000 ×g for 5 min at RT, the cell pellets were washed twice with PBS straightforward to accomplish, previous attempts to increase CL levels buffer, snap frozen in liquid N2, and stored in a – 80 °C freezer until were not successful by direct administration of this lipid in vivo [29]. used. Hence, it was hypothesized that by promoting plasmalogen biosynthesis it would be possible to restore to normal the levels of this lipid in Barth specimens and, eventually, normalize the lipidome and mitochondria 2.3. Human lymphoblast lipid extraction for high resolution 31P NMR function. To do so lymphoblasts derived from BTHS patients were used and compared with controls. Cells were administered with plasmalogen Lipid extraction from lymphoblast cells for high resolution 31P NMR precursors and the lipid profile was characterized by high resolution 31P was conducted as described in Kimura and collaborators [21]. Briefly, a NMR phospholipidomic analysis and mass spectrometry and compared Cs-EDTA buffer was prepared by titration of a 200 mM EDTA solution with cells without the administration of plasmalogen precursors. The with 50 wt% aqueous CsOH until pH 6.0 was reached, addition of effect of the treatment on cell viability, mitochondria biogenesis, and 50 μM BHT, readjustment of the pH to 6.0, and degassing by argon mitochondria membrane potential were also evaluated. bubbling [31]. Lymphoblast cell pellets were thawed on ice, re- It is shown that the promotion of plasmalogen biosynthesis could suspended in cold Cs-EDTA buffer (1.2 mL of buffer to ca. 600 μL cell restore the low levels of PE-Pls in this BTHS cell model to levels com- pellet), and homogenized by striking the cell suspension 30 times with a parable to the control. In addition, the data showed a compensatory Teflon pestle in a tight-fit glass vessel. Lipids from the homogenized cell effect of the diacyl counterpart lipid upon plasmalogen precursors ad- suspension were, then, extracted by the Folch method using a chlor- ministration; that is, a decrease in the steady-state levels of diacyl-PE. oform:methanol (2,1, vol/vol) solvent mixture containing 250 μM BHT Noteworthy, promotion of plasmalogen biosynthesis led to a significant [32]. After removal of the top aqueous phase, the organic phase was increase in the levels of CL in BTHS lymphoblasts, while no changes washed twice with cold Cs-EDTA buffer. The organic phase was, then, were observed in the MLCL levels. Promotion of plasmalogen bio- collected, dried under a N2 flux, and vacuum-dried for at least 2 h to synthesis did not impact cell viability, although it significantly decrease form a lipid film. For high resolution 31P NMR measurements, lipid mitochondria copy number and restored mitochondrial membrane po- films were hydrated with 600 μL of degassed 10% (wt/v) SDS buffer tential to levels comparable to controls. Overall, the results highlight (50 mM MES, 50 μM BHT, pH 6.0) containing 10% (v/v) D2O.

2 J.C. Bozelli, et al. BBA - Molecular and Cell Biology of Lipids 1865 (2020) 158677

2.4. High-resolution 31P NMR Briefly, 50 μL of lysis buffer were added to 100 μL of a lymphoblast (at a cell density of ca. 106/mL) cell suspension per well in a 96-well mi- The phospholipidomic analysis of the lipids extracted from the croplate and incubated for 5 min at RT in a rocking shaker. Following, lymphoblasts was conducted by measuring the high-resolution 31P NMR 50 μL of luciferase/d-luciferin solution were added to each well and spectra of the SDS solubilized lipids as described by Kimura et al. [21]. incubated for 5 min at RT in a rocking shaker. The plate was dark Briefly, samples were placed in 5-mm-diameter NMR tubes and their adapted for 10 min and the luminescence was measured in a Synergy 4 spectra were acquired on a Bruker AVANCE-III 700 MHz spectrometer microplate reader (BioTek, Winooski, VT). An ATP standard solution (31P frequency, 283.4 MHz) equipped with a QNP cryoprobe at 25 °C. was used to build a calibration curve and the number of cells was Spectra were acquired with an 80° excitation pulse on 31P nuclei and quantified in a Countless cell counter from Invitrogen (Carlsbad, CA). inverse-gated broadband 1H decoupling with the WALTZ-16 sequence at a 3.8 kHz decoupling frequency. Spectra were acquired over 12 ppm 2.7. Mitochondria copy number (3.4 kHz) bandwidth, 2.4 s acquisition time, 1.0 s recycle delay, and 2048 scans. The chemical shift scale was referenced to an external 85% The mitochondria copy number in the lymphoblast cells were (wt/v) phosphoric acid standard set to 0 ppm. measured by quantifying the ratio of mitochondrial to nuclear DNA as previously described [34]. Briefly, the genomic DNA of lymphoblasts 2.5. Human lymphoblast lipid extraction and LC-MS analysis was extracted using NucleoSpin Tissue kit from Takara Bio (Japan) and had their concentration and purity quantified using a nanodrop. In a Lipid extractions were performed as previously described [33]. real-time PCR 96-well microplate (BioRad – Hercules, CA) the real time Briefly, lymphoblast cell pellets were resuspended in 1 mL of cold PBS. PCR master mix (TB green premix Ex TaqII, Tli RNAseH Plus; Takara An aliquot of the mixture was taken to measure protein concentration Bio – Japan) was added, followed by a set of four optimized PCR primer using a Coomassie (Bradford) Protein Assay Kit (Thermo Scientific, pairs (one per well) targeting two nuclear and two mitochondrial genes Waltham, MA) following the manufacturer's instructions. Lipids were from Takara Bio (Japan) and 10 ng of genomic DNA. Finally, real-time extracted after adding in 1 mL of methanol and 2 mL of chloroform into PCR was performed in a CFX96 qPCR from BioRad (Hercules, CA) and the remaining material in PBS using a glass Dounce homogenizer. The the ratio between mitochondrial and nuclear DNA was determined from sample was transferred into a 2 dram glass vial and was centrifuged at the differences in Ct values for mitochondrial and nuclear DNA. 4 °C, 500 ×g for 15 min to separate the aqueous and organic phases. The organic phase was transferred into a 1 dram glass vial and the 2.8. Mitochondria membrane potential solvent was removed by rotary evaporation. The dried lipids were resuspended in chloroform, using volumes calculated to normalize sam- The mitochondria membrane potential was measured as previously ples based on protein concentration. Lipids were analyzed by LC-MS described [11]. Briefly, ca. 106 cells were incubated in 1.5 mL culture using Agilent Infinity 1260 HPLC / Agilent 6530 Jet Stream ESI-QTof- medium with 7.5 μM JC-1 (a mitochondrial membrane potential probe) MS system using methods that were previously described [33]. The from EMD Millipore (Burlington, MA) for 10 min at 37 °C in a humi- lipid extracts for each condition were combined and aliquoted for dified 5% CO2 atmosphere. Stained cells were washed three times with analysis. PBS and suspended in 1 mL PBS containing 5% (g/100 mL) BSA. Fol- For analysis of diacyl PE species, Mobile phase A consisted of 95% lowing, stained cells were filtered with a Falcon round-bottom tube water and 5% methanol, Mobile phase B consisted of 60% isopropanol, with a cell strainer and immediately analyzed by flow cytometry in a 35% methanol and 5% water in negative ionization mode. The se- BD LSR Fortessa™ flow cytometer and data were analyzed by FlowJo (v. paration was performed using a Gemini C18 column with a C18 guard 10.6.1) software. The aggregate/monomer JC-1 emission ratio (red/ cartridge, and 0.1% (w/v) ammonium hydroxide was added to the green fluorescence ratio) was used as an estimate of mitochondrial mobile phases. The analysis of samples was performed using 5 μL in- membrane potential. As a reference, uncoupled mitochondria (cells jection volume and began with 5 min of 0% B at 0.1 mL/min, then the pretreated with 1 μM FCCP) were used. Data is presented as the ap- flow rate was increased to 0.5 mL/min and a mobile phase gradient parent mitochondrial membrane potential (ΔΨApp Red Green Red Green Red Green began with 0% B to 100% B over 60 min, maintained at 100% at B for (%) = [(F :F − FFCCP:FFCCP )/F :F ] × 100). 7 min and switched to 0% B for 8 min. Data analysis was performed using Agilent MassHunter Qualitative Analysis Software where corre- 2.9. Statistical analysis sponding m/z's for each ion were extracted and peak areas for each ion were manually integrated. Statistical analysis was done by performing one-way ANOVA test on For analysis of MLCL and CL species, Mobile phase A used consist of data set and statistical significance was assumed for p-values < 0.05. 95% water and 5% methanol, Mobile phase B consisted of 60% isopropanol, 35% methanol and 5% water in negative ionization mode. 3. Results The separation was performed using a Luna C5 column with a C5 guard cartridge, and 0.1% (v/v) ammonium acetate was added to the mobile 3.1. Effect of the promotion of plasmalogen biosynthesis on the phases. The analysis of samples was performed using 5 μL injection phospholipidome of BTHS lymphoblasts volume and began with 5 min of 0% B at 0.1 mL/min, then the flow rate was increased to 0.5 mL/min and a mobile phase gradient began with To reverse the low levels of plasmalogens found in Barth specimens, 0% B to 100% B over 40 min, maintained at 100% at B for 7 min and a strategy based on the promotion of their biosynthesis was used. The switched to 0% B for 8 min. Chromatographic alignments and data effects of the promotion of plasmalogen biosynthesis were studied using analysis were carried out using Agilent MassHunter Profinder Software lymphoblasts derived from BTHS patients and compared with controls. due to the low abundances of these species. The cells from Barth patients have all the characteristics of cells from patients with this Syndrome, including the decrease in the levels of 2.6. Cellular ATP content plasmalogens [9,11,21]. Plasmalogen biosynthesis begins in peroxi- somes and is finalized in the endoplasmic reticulum (ER) where mature Cell viability was estimated by measuring the ATP content of the plasmalogen is formed and prepared to be transported to plasma and lymphoblasts using the ATPlite (Perkin-Elmer, Waltham, MA) assay organelle membranes [22]. 1-O-alkylglycerols are plasmalogen pre- system, which is based on the production of light (luminescence) cursors that enter the biosynthesis pathway at the ER; i.e., downstream caused by the reaction of ATP with added luciferase and d-luciferin. of peroxisomal steps [22]. Administration of alkylglycerols in vitro and

3 J.C. Bozelli, et al. BBA - Molecular and Cell Biology of Lipids 1865 (2020) 158677

Fig. 1. High resolution 31P NMR spectra of lymphoblasts phospholipids. (A) High resolution 31P NMR spectra of phospholipids extracted from control, C, and Barth, BTHS, lymphoblasts are overlayed for comparison. (B) High resolution 31P NMR spectra of phospholipids extracted from Barth lymphoblasts, BTHS, and Barth lymphoblasts adminis- ⁎ tered with plasmalogen precursors, BTHS , are overlayed for comparison. The insert shows a mag- niﬁcation of the spectral region highlighted in red. Spectra are representative of 4 independent measurements. Peaks are numbered and the legend presents the assignment of each peak to a phospholipid class (or subclass). PC: diacyl-phosphatidylcholine (+plasmanyl-phosphatidylcholine), PI: phosphatidylinositol, PS: phosphatidylserine, LPC: lyso-phosphatidylcholine, LPC-Pls: lyso-plasmenyl- phosphocholine, PE diacyl-phosphatidylethanolamine (+plasmanyl-phosphatidylethanolamine), PE- Pls: plasmenyl-phosphoethanolamine, SM: sphingo- myelin, GPC: glycerophosphocholine, LPS: lyso- phosphoserine, LPE: lyso-phosphoethanolamine, LPE-Pls: lyso-plasmenylphosphoethanolamine, CL: cardiolipin, PG: phosphatidylglycerol, 1-MLCL: 1- monolyso-cardiolipin, and 2-MLCL: 2-monolysocardiolipin.

in vivo has shown to be a successful strategy to increase plasmalogen Table 1 levels [28,35–39]. Hence, BTHS and control lymphoblasts were ad- Effect of promotion of plasmalogen biosynthesis on the phospholipidome of ministered with an alkylglycerol, namely 1-O-hexadecylglycerol (HG), Barth lymphoblasts. Phospholipids, their 31P NMR chemical shifts, and in- aiming to promote plasmalogen biosynthesis. In addition to HG, etha- dividual contents (mol%, relative to total phospholipid) are presented. Values are presented for control (C) and Barth (BTHS) lymphoblasts as well as Barth nolamine was co-administrated to avoid it being a rate limiting factor as ⁎ the major plasmalogen in lymphoblasts is plasmenylethanolamine (PE- (BTHS ) lymphoblasts that have been administered with plasmalogen precursors. Data were generated from the simulation of the high resolution 31P Pls) [21,30]. NMR spectra presented in Fig. 1 and are presented as mean ± S.D. (n = 4). To evaluate the effects of promotion of plasmalogen biosynthesis on the phospholipidome of the lymphoblasts, high resolution 31P NMR Lipid Chemical shift (ppm) C BTHS BTHS* spectra were measured on their lipid extracts. This methodology pre- PCa −0.174 54 ± 2 52 ± 2 49.1 ± 0.5c,d sents high selectivity as well as ease of performance and analysis for PI 0.033 5.2 ± 0.3 4.6 ± 0.8 6.0 ± 0.5d lipid extracts. Sharp NMR peaks are obtained by solubilizing the lipids PS 0.114 4.8 ± 0.2 4.3 ± 0.4 4.8 ± 0.4 in detergent micelles (SDS in the present study), which leads to unique LPC 0.204 0.3 ± 0.2 0.04 ± 0.03 0.3 ± 0.1d b b chemical shift values of the phosphate group for the different phos- LPC-Pls 0.220 n/d n/d 0.15 ± 0.02 PEa 0.284 16 ± 1 23 ± 1c 13.1 ± 0.8c,d pholipid classes and subclasses avoiding the need of separation of PE-Pls 0.324 11.5 ± 0.6 9.0 ± 0.5c 17.9 ± 0.7c,d components of the mixture prior to measurement. Fig. 1A presents the SM 0.404 1.9 ± 0.5 2.7 ± 0.3c 2.6 ± 0.1c overlapped high resolution 31P NMR spectra of BTHS and control GPC? 0.466 n/db 0.3 ± 0.2 0.4 ± 0.3 lymphoblast lipid extracts. These spectra are characterized by the LPS 0.513 1.4 ± 0.8 0.5 ± 0.4 0.9 ± 0.6 presence of 12–14 peaks, which are ascribed to different phospholipid LPE 0.662 0.3 ± 0.2 0.13 ± 0.06 0.07 ± 0.01 LPE-Pls 0.721 n/db n/db 0.6 ± 0.3 classes and subclasses. Kimura and collaborators have previously as- CL 0.736 2.1 ± 0.2 0.9 ± 0.3c 1.44 ± 0.09c,d signed most of these peaks based on comparison with spectra of phos- PG 0.838 0.8 ± 0.2 1.6 ± 0.3c 1.20 ± 0.06c,d pholipid standards, literature data, and the use of different detergents 1-MLCL 1.009 n/db 0.09 ± 0.04 0.08 ± 0.01 b [20,21]. Table 1 summarizes the relative abundance of the individual 2-MLCL 1.131 n/d 0.5 ± 0.2 0.7 ± 0.2 phospholipids identified in lymphoblasts of BTHS and control in- a These phospholipids comprise the sum of both their diacyl and plasmanyl ff dividuals. The results showed di erences in the content of 6 phospho- molecular species. lipid species; those were PE, PE-Pls, SM, CL, PG, and MLCL. In BTHS b These lipid species were not detected (n/d). lymphoblasts the content of PE, SM, PG, and MLCL were significantly c Statistical significant in comparison to C (p < 0.05). increased, while that of PE-Pls and CL were decreased, in agreement d Statistical significant in comparison to BTHS (p < 0.05). with the previous report [21]. On average PE, SM, and PG increased their individual content by 44, 42, and 100%, respectively; and PE-Pls high resolution 31P NMR. and CL decreased their individual content by 22 and 57%, respectively Upon administering lymphoblasts with plasmalogen precursors no- (Fig. 2 and Table 1). MLCL was not detected in control lymphoblasts by table changes were observed in the high resolution 31P NMR spectra of

4 J.C. Bozelli, et al. BBA - Molecular and Cell Biology of Lipids 1865 (2020) 158677

Fig. 2. Eﬀect of promotion of plasmalogen biosynthesis on the individual content of selected phospholipids of Barth lymphoblasts. Data was generated from the normalization of their individual content presented in Table 1 to their individual abundance in controls, C, and are presented for (A) PE-Pls, (B) diacyl-PE (+ plasmanyl-PE), (C) CL, and (D) 2-MLCL. Data is presented as a box plot (n = 3 ⁎ for C and 4 for BTHS and BTHS ; *p < 0.05). Note that since there was no MLCL measured in controls, in (D) data was normalized for the values in Barth lymphoblasts, BTHS,. Data is only presented for 2- MLCL as this was the isomer with higher content, although a similar trend was observed for 1-MLCL ⁎ (not shown). BTHS are the values for Barth lymphoblasts that were administered with plasmalogen precursors. Values reﬂect data obtained by high resolution 31P NMR measurements.

their lipid extracts (Figs. 1B and S1). A summary of the relative abun- shorter less unsaturated acyl chains (34:1 and 36:2) were affected dance of the individual phospholipids identified in BTHS lymphoblasts modestly. CL molecular species that were detected mainly showed de- that were administered with plasmalogen precursors is presented in pletions in BHTS as compared to control lymphoblasts (72:6, 72:7, 72:8 Table 1. The phospholipid species that were significantly affected by and 72:9, Fig. 3B), except for 72:0 and 72:2 molecules species which administering plasmalogen precursors to BTHS lymphoblasts were PE, were less affected. In control lymphoblasts two MLCL species were PE-Pls, CL, and PG. In addition, two lysophospholipids were identified detected, albeit their low abundances (Fig. 3C). 54:5 MLCL was de- in BTHS lymphoblasts that were administered with plasmalogen pre- tected at similar levels in control and BHTS; and 54:6 showed a slight cursors, namely, LPC-Pls and LPE-Pls (Fig. 1B and Table 1). When accumulation (Fig. 3C). compared with BTHS lymphoblasts, administering plasmalogen pre- Upon promotion of plasmalogen biosynthesis, LC-MS data showed cursors led, on average, to 43 and 25% decrease in the individual that the majority of the PE molecular species detected were depleted in content of PE and PG, respectively; and a 100 and 60% increase in the BTHS lymphoblasts (BTHS vs BTHS*, Fig. 3A), except for longer poly- individual content of PE-Pls and CL, respectively (Fig. 2 and Table 1). unsaturated acyl chains (36:4, 38:3 and 38:4). The results also showed Interesting, in control lymphoblasts administered with plasmalogen that the CL molecular species profile is not significantly changed upon precursors, a similar trend in the relative change of these lipid species the administration of plasmalogen precursors (BTHS and BTHS*), ex- as well the presence of LPC-Pls and LPE-Pls were observed, corrobor- cept for C72:0 CL which showed a modest accumulation in BTHS* when ating the finding in BTHS lymphoblasts administered with plasmalogen compared to BTHS (Fig. 3B). The promotion of plasmalogen biosynth- precursors (Figs. S1–2 and Table S1). It is important to mention though esis did not change the MLCL molecular species profile when comparing that none of these changes restored their individual content found in BTHS lymphoblasts with or without administered plasmalogen pre- control lymphoblasts (Figs. 2 and S2 and Tables 1 and S1). In addition, cursors (Fig. 3C). In control lymphoblasts administered with plasma- as MLCL accumulation is a hallmark of BTHS it is important to mention logen precursors, the majority of the PE and MLCL species detected that administration of plasmalogen precursors to BTHS lymphoblasts were depleted; whereas there was an accumulation of the majority of did not change the levels of this lipid (Fig. 2D). To deepen the under- the CL species detected in comparison to controls, supporting the standing of the observed phospholipid changes upon promoting plas- findings obtained by 31P NMR (Fig. S4). malogen biosynthesis in these lymphoblasts liquid chromatography- mass spectrometry (LC-MS) was conducted targeting three phospholi- 3.2. Effect of the promotion of plasmalogen biosynthesis on cell viability and pids; namely, PE, CL, and MLCL (Figs. 3 and S4). In control lympho- mitochondrial properties of BTHS lymphoblasts blasts the major PE molecular species bears 36:1 acyl chains, which is also the prevalent molecular species in BTHS lymphoblasts (Fig. 3A). Overall, these results indicate the success of the strategy to restore The major change in the PE molecular species profile in BTHS lym- the low levels of PE-Pls in BTHS lymphoblasts. Of note, was the ob- phoblasts when compared to control was an overall increase in longer servation that this strategy also led to a significant increase in the levels polyunsaturated acyl chains (36:4, 38:3 and 38:4) (Fig. 3A) whereas of CL. However, to substantiate these findings having the perspective of

5 J.C. Bozelli, et al. BBA - Molecular and Cell Biology of Lipids 1865 (2020) 158677

Fig. 3. Effect of promotion of plasmalogen biosynthesis on the molecular species profile of targeted phospholipids of Barth lymphoblasts. Heatmap representation of relative abundance of different molecular species of (A) PE, (B) CL, and (C) MLCL are shown. Relative abundances were calculated by dividing the raw abundance of each lipid molecular species by the average abundance of the same lipid species in all samples. Red coloring indicates high levels while blue indicates low levels. Species that were not detected were assigned zero (0) abundance. Each sample in the heatmaps displays the values of 4 independent measurements. Valuesreflect data obtained by mass spectrometry measurements. developing a strategy to alleviate/treat BTHS patients, it's important to The measure of mitochondrial membrane potential is an indicator of evaluate the effect of the promotion of plasmalogen biosynthesis on cell respiratory chain maturation [40]. It has been shown that in Barth viability. In addition, since BTHS mitochondria are dysfunctional it is specimens mitochondrial membrane potential is lower than in control also relevant to evaluate the effect of administering the cells plasma- cells, which is a consequence of a lower content and more hetero- logen precursors on mitochondria biogenesis and functional properties. geneous CL molecular species decreasing the stability of respiratory To evaluate the effect of administering the lymphoblasts plasmalogen complexes and SC [11]. To evaluate the effect of the promotion of precursors on cellular viability the measurement of cellular ATP was plasmalogen biosynthesis on the mitochondrial membrane potential we used as an estimation of cell proliferation and cytotoxicity. Fig. S3 measured the aggregate/monomer (red/green) JC-1 fluorescence ratio compares the total cellular ATP content of control and BTHS lympho- as an estimation of mitochondrial membrane potential (Figs. 4B and blasts with and without the administration of plasmalogen precursors. S5B). BTHS lymphoblasts showed, on average, a 34% lower apparent BTHS lymphoblasts had ca. 2.3-fold higher total cellular ATP than mitochondrial membrane potential (Fig. 4B). Interestingly, upon pro- controls, which is in agreement with the known cellular compensatory motion of plasmalogen biosynthesis the mitochondrial membrane po- effects of BTHS lymphoblasts to circumvent the disrupted mitochon- tential was fully restored to levels compared to control lymphoblasts dria; i.e., higher glycolysis and mitochondria content [9,11,18]. Plas- (Fig. 4B). This result was also corroborated by a small, but significant, malogen precursors' administration didn't change significantly the cel- 13% average increase in the apparent mitochondrial membrane po- lular ATP content of BTHS lymphoblasts, indicating that the treatment tential of control lymphoblasts that were administered plasmalogen did not impact to a large extent on cell proliferation and cytotoxicity precursors (Fig. S5B). (Fig. S3). No changes were also observed in the cellular ATP content of control lymphoblasts that were administered plasmalogen precursors, 4. Discussion corroborating the indication that administering lymphoblasts with plasmalogen precursors did not impact on cell proliferation and/or Currently, there is no specific treatment for BTHS. Contrary to the cytotoxicity (Fig. S3). One of the strategies BTHS cells use to compen- well-established alterations in CL content and molecular species in sate their mitochondrial dysfunction is to increase mitochondria bio- BTHS, the recent identification of a marked decrease in another class of genesis; i.e., its mitochondrial copy number [9]. To evaluate the effect lipids, namely plasmalogens, in BTHS has opened a new possibility, up of the promotion of plasmalogen biosynthesis on the mitochondrial to now not considered, as a potential strategy to alleviate/treat this copy number we measured the ratio of mitochondrial to nuclear DNA in condition, which is through the promotion of plasmalogen biosynthesis the lymphoblasts, which is an accurate way to quantitate the mi- [20,21]. It has been shown in other systems that restoring plasmalogen tochondria copy number (Figs. 4A and S5A). The results showed that levels is beneficial at the cellular and mitochondrial levels, in restoring BTHS lymphoblasts had on average 3-fold more mitochondria than biological function [22,28]. Noteworthy, in several diseases char- control lymphoblasts, in agreement with the literature [9,11]. Inter- acterized by depletion in plasmalogens the use of plasmalogen re- estingly, upon administering plasmalogen precursors to BTHS lym- placement therapy (via the promotion of their synthesis) has been phoblasts there was a small but significant decrease in the mitochondria shown as a successful strategy to alleviate/treat those conditions [22]. copy number, which now had, on average, 1.8-fold more mitochondria Hence, the present work describes the effect of the promotion of plas- than control lymphoblasts, a 40% decrease from BTHS lymphoblasts malogen biosynthesis on the phospholipidome of BTHS lymphoblasts, (Fig. 4A). However, it should be mentioned that contrary to BTHS which is a simple cell model that bears all the characteristics of this lymphoblasts, in control lymphoblasts administering plasmalogen pre- disease. In addition, an evaluation of the administration of the plas- cursors led to an increase in the mitochondria copy number, which was, malogen precursors on cell viability, mitochondria biogenesis, and on average, 1.8-fold higher than control lymphoblasts (Fig. S5A). mitochondrial membrane potential were evaluated. It is shown that

6 J.C. Bozelli, et al. BBA - Molecular and Cell Biology of Lipids 1865 (2020) 158677

Fig. 4. Eﬀect of promotion of plasmalogen biosynthesis on mitochondrial properties of Barth lymphoblasts. (A) Mitochondria copy number and (B) Apparent mitochondrial membrane potential. Data were normalized for values of control lymphoblasts, C. Data are presented as a box plot (n = 6 and 3 in A and B, respectively; *p < 0.05) for Barth lymphoblasts without administration (BTHS) and with ad- ⁎ ministration (BTHS ) of plasmalogen precursors.

administering BTHS lymphoblasts with plasmalogen precursors led to The hallmark of BTHS is the alterations that happen in CL content an increase in the low levels of PE-Pls (the main plasmalogen in these and molecular species due to the lack of functional tafazzin [5,7,45]. cells) with a counterbalancing decrease in its diacyl-PE content. Of note Previous attempts to increase CL levels were not successful by direct is the significant increase in the content of CL, while not altering the administration of this lipid in vivo [29]. Interestingly, the results pre- content of its hydrolysis product MLCL. Moreover, the treatment did sented here showed that by promoting plasmalogen biosynthesis in not impact on cellular viability, although it significantly decrease mi- addition to increasing plasmalogen levels, it also leads to an increase in tochondria copy number and restored mitochondrial membrane po- CL levels. This result suggests that increased plasmalogen levels might tential. regulate CL biosynthesis and/or degradation pathways. The under- The total cellular level of PE glycerophospholipids tends to remain standing of the molecular mechanisms that triggered CL content in- constant. It has been previously shown that to keep the overall PE crease upon promotion of plasmalogen biosynthesis is beyond the scope content constant, a decrease in PE-Pls content is compensated with an of the present work. However, it seems reasonable to propose that the increase in the content of its diacyl-PE counterpart; likewise, an in- regulation might be happening in either one or a combination of the crease in PE-Pls content is compensated with a decrease in the content following biochemical reactions that control CL levels; those are: (i) CL of its diacyl-PE counterpart [41–43]. The total cellular content of PE hydrolysis by a phospholipase (iPLA2β, iPLA2γ, cPLA2, and/or sPLA2), lipids derived from the Kennedy pathway are controlled by the levels of (ii) MLCL acylation by ALCAT1 and/or MLCLAT1, (iii) protection of CL CDP-ethanolamine, through the action of CTP:phosphoethanolamine oxidation, and (iv) increased CL synthesis [46–48]. The results pre- cytidyltransferase. CDP-ethanolamine is coupled to alkylacylglycerol or sented here suggest that (i) and (ii) are less likely than (iii) and (iv) if diacylglycerol to yield PE-Pls and diacyl-PE, respectively [43]. Hence, if those are considered as single rather a combination of reactions affected the level of CDP-ethanolamine is kept constant, the relative amounts of by increased plasmalogen levels. The reason is that in (i) and (ii) by PE-Pls and diacyl-PE will adjust according to the relative levels of al- themselves a decrease in the steady-state levels of MLCL would be ex- kylacylglycerol and diacylglycerol, respectively. The studies reported pected upon promotion of plasmalogen biosynthesis, which was not here are in very good agreement with previous findings showing that observed. Future studies will shed some light on how promotion of the total PE content of lymphoblasts are held constant at ca. 30 mol% plasmalogen biosynthesis leads to an increase in CL content. [21]. In addition, the results suggested that the low PE-Pls:diacyl-PE In BTHS the alterations in CL as a consequence of the lack of ta- ratio found in BTHS lymphoblasts are due to a decreased alkyla- fazzin function lead to dysfunctional mitochondria [9,11,12,18]. The cylglycerol:diacylglycerol ratio. This is corroborated by the present results presented here show that upon promotion of plasmalogen bio- studies where by increasing the alkylacylglycerol:diacylglycerol ratio synthesis mitochondria copy number was decreased and mitochondrial by administering lymphoblasts with plasmalogen precursors led to an membrane potential was fully restored to control levels. It is known that increase in the levels of PE-Pls with a concomitant decrease in the levels in BTHS cellular compensation occurs due to the dysfunctional mi- of diacyl-PE. tochondria, one of which is the increase in mitochondria copy number Plasmalogens are usually enriched with PUFAs at the sn-2 position as a consequence of a less robust electron transport chain activity [9]. of the glycerol moiety [22]. PUFAs are biologically active mediators of The decreased in the mitochondria copy number upon promoting signaling processes, which cannot be synthesized de novo in higher plasmalogen biosynthesis suggested that the treatment led to a more mammals and need to be acquired through the diet; i.e., they are “es- robust electron transport chain activity. Indeed, the finding that the sential fatty acids”. Thus, their levels are tightly regulated. It has been mitochondrial membrane potential was fully restored upon promotion shown that in conditions where plasmalogen levels are decreased, the of plasmalogen biosynthesis corroborated this indication. It is known increase in diacyl-PE is favored for those molecular species enriched in that plasmalogen imparts different physical properties to the membrane PUFAs, which seems to be a way to maintain the total PUFAs levels in when compared to their diacyl counterparts [23]. Thus, it is con- these lipids [43,44]. The results presented here are in very good ceivable that the increase in plasmalogen content had led to a more agreement with the literature, showing that in BTHS lymphoblasts the packed membrane, which would, in turn, make the membrane less counterbalanced increase in diacyl-PE due to the decreased PE-Pls le- “leaky”, increasing the mitochondria membrane potential. Likewise, vels are preferential for those diacyl-PE molecular species enriched in the increase in CL levels could have led to a more efficient electron PUFAs. Interestingly, the promotion of plasmalogen biosynthesis, al- transport chain activity. though it increased the total cellular levels of PE-Pls with a compensatory decrease in the total cellular levels of its diacyl-PE counterpart, it did not change the profile of the diacyl-PE molecular species, which 5. Conclusions continued to be enriched in PUFAs. To sum up, the present study showed that it is possible to restore the

7 J.C. Bozelli, et al. BBA - Molecular and Cell Biology of Lipids 1865 (2020) 158677 low levels of plasmalogen in a BTHS cell model upon promotion of its Petit, Barth syndrome: cellular compensation of mitochondrial dysfunction and biosynthesis. Importantly, the administration of plasmalogen pre- apoptosis inhibition due to changes in cardiolipin remodeling linked to tafazzin (TAZ) gene mutation, Biochim. Biophys. Acta - Mol. Basis Dis. 1832 (2013) cursors did not impact on the cell viability, while it increased to some 1194–1206. doi:https://doi.org/10.1016/j.bbadis.2013.03.005. extent the CL levels as well as restored mitochondrial properties. Until [10] N. Ikon, R.O. Ryan, Barth syndrome: connecting cardiolipin to cardiomyopathy, the association of plasmalogen loss with BTHS such strategy could not Lipids 52 (2017) 99–108, https://doi.org/10.1007/s11745-016-4229-7. [11] Y. Xu, J.J. Sutachan, H. Plesken, R.I. Kelley, M. Schlame, Characterization of have been considered and the results presented here suggest that a diet lymphoblast mitochondria from patients with Barth syndrome, Lab. Investig. 85 supplemented with plasmalogen precursors could be beneficial for (2005) 823–830, https://doi.org/10.1038/labinvest.3700274. BTHS patients. [12] G. Wang, M.L. McCain, L. Yang, A. He, F.S. Pasqualini, A. Agarwal, H. Yuan, D. Jiang, D. Zhang, L. Zangi, J. Geva, A.E. Roberts, Q. Ma, J. Ding, J. Chen, D.Z. Wang, K. Li, J. Wang, R.J.A. Wanders, W. Kulik, F.M. Vaz, M.A. Laflamme, Declaration of competing interest C.E. Murry, K.R. Chien, R.I. Kelley, G.M. Church, K.K. Parker, W.T. Pu, Modeling the mitochondrial cardiomyopathy of Barth syndrome with induced pluripotent – The authors declare that they have no conflicts of interest with the stem cell and heart-on-chip technologies, Nat. Med. 20 (2014) 616 623, https:// doi.org/10.1038/nm.3545. content of this article. [13] M. McKenzie, M. Lazarou, D.R. Thorburn, M.T. Ryan, Mitochondrial respiratory chain supercomplexes are destabilized in Barth syndrome patients, J. Mol. Biol. 361 – Acknowledgements (2006) 462 469, https://doi.org/10.1016/j.jmb.2006.06.057. [14] Y. Xu, C.K.L. Phoon, B. Berno, K. D’Souza, E. Hoedt, G. Zhang, T.A. Neubert, R.M. Epand, M. Ren, M. Schlame, Loss of protein association causes cardiolipin We appreciate the technical support of Hong Liang (McMaster degradation in Barth syndrome, Nat. Chem. Biol. 12 (2016) 641–647, https://doi. University) with flow cytometry measurements and Omar El-Halfawy org/10.1038/nchembio.2113. [15] Y. Xu, M. Anjaneyulu, A. Donelian, W. Yu, M.L. Greenberg, M. Ren, E. Owusu- (McMaster University) with real-time PCR measurements. We thank Ansah, M. Schlame, Assembly of the complexes of oxidative phosphorylation trig- Shamin Fernando for his help with lipid extraction for LC-MS analysis. gers the remodeling of cardiolipin, Proc. Natl. Acad. Sci. U. S. A. 116 (2019) 11235–11240, https://doi.org/10.1073/pnas.1900890116. [16] F. Diaz, J.A. Enriquez, C.T. Moraes, Cells lacking Rieske iron-sulfur protein have a Funding reactive oxygen species-associated decrease in respiratory complexes I and IV, Mol. Cell. Biol. 32 (2012) 415–429, https://doi.org/10.1128/mcb.06051-11. This work was supported by the Natural Sciences and Engineering [17] E. Maranzana, G. Barbero, A.I. Falasca, G. Lenaz, M.L. Genova, Mitochondrial re- Research Council of Canada (NSERC) grant RGPIN-2018-05585 to spiratory supercomplex association limits production of reactive oxygen species from complex I, Antioxid. Redox Signal. 19 (2013) 1469–1480, https://doi.org/10. R.M.E and the National Science Foundation grant MCB1817468 to 1089/ars.2012.4845. G.E.A.-G. and the Research Foundation for the State University of New [18] E.M. Mejia, H. Zegallai, E.D. Bouchard, V. Banerji, A. Ravandi, G.M. Hatch, York (J. Solo funds) to G.E.A.-G.. Expression of human monolysocardiolipin acyltransferase-1 improves mitochondrial function in Barth syndrome lymphoblasts, J. Biol. Chem. 293 (2018) 7564–7577, https://doi.org/10.1074/jbc.RA117.001024. Author contributions [19] S. Chen, Q. He, M.L. Greenberg, Loss of tafazzin in yeast leads to increased oxidative stress during respiratory growth, Mol. Microbiol. 68 (2008) 1061–1072, https:// fi doi.org/10.1111/j.1365-2958.2008.06216.x. J.C.B. Jr. performed cell work, cellular ATP content quanti cation, [20] T. Kimura, A.K. Kimura, M. Ren, B. Berno, Y. Xu, M. Schlame, R.M. Epand, 31 mitochondria copy number quantification, and high resolution P Substantial decrease in plasmalogen in the heart associated with tafazzin defi- NMR measurements. D.L. performed mass spectrometry measurements ciency, Biochemistry. 57 (2018) 2162–2175, https://doi.org/10.1021/acs.biochem. 8b00042. under the supervision of G.E.A.-G. J.C.B. Jr. and R.M.E. designed the [21] T. Kimura, A.K. Kimura, M. Ren, V. Monteiro, Y. Xu, B. Berno, M. Schlame, R.M. experiments, analyzed the data, and wrote the manuscript. Epand, Plasmalogen loss caused by remodeling deficiency in mitochondria, Life Sci. Alliance. 2 (2019) pii: e201900348. doi:10.26508/lsa.201900348. [22] N.E. Braverman, A.B. Moser, Functions of plasmalogen lipids in health and disease, Appendix A. Supplementary data Biochim. Biophys. Acta, Mol. Basis Dis. 1822 (2012) 1442–1452, https://doi.org/ 10.1016/j.bbadis.2012.05.008. Supplementary data to this article can be found online at https:// [23] X. Han, R.W. Gross, Plasmenylcholine and phosphatidylcholine membrane bilayers – doi.org/10.1016/j.bbalip.2020.158677. possess distinct conformational motifs, Biochemistry 29 (1990) 4992 4996, https://doi.org/10.1021/bi00472a032. [24] F. Paltauf, Ether lipids in biomembranes, Chem. Phys. Lipids 74 (1994) 101–139, References https://doi.org/10.1016/0009-3084(94)90054-X. [25] D. Reiss, K. Beyer, B. Engelmann, Delayed oxidative degradation of polyunsaturated diacyl phospholipids in the presence of plasmalogen phospholipids in vitro, [1] H.B. Neustein, P.R. Lurie, B. Dahms, M. Takahashi, An X-linked recessive cardio- Biochem. J. 323 (1997) 807–814, https://doi.org/10.1042/bj3230807. – myopathy with abnormal mitochondria, Pediatrics 64 (1979) 24 29. [26] B. Engelmann, C. Bräutigam, J. Thiery, Plasmalogen phospholipids as potential ’ [2] S. Bione, P. D Adamo, E. Maestrini, A.K. Gedeon, P.A. Bolhuis, D. Toniolo, A novel protectors against lipid peroxidation of low density lipoproteins, Biochem. Biophys. X-linked gene, G4.5. is responsible for Barth syndrome, Nat. Genet. 12 (1996) Res. Commun. 204 (1994) 1235–1242, https://doi.org/10.1006/bbrc.1994.2595. – 385 389, https://doi.org/10.1038/ng0496-385. [27] D. Hahnel, K. Beyer, B. Engelmann, Inhibition of peroxyl radical-mediated lipid [3] Y. Xu, A. Malhotra, M. Ren, M. Schlame, The enzymatic function of tafazzin, J. Biol. oxidation by plasmalogen phospholipids and α-tocopherol, Free Radic. Biol. Med. – Chem. 281 (2006) 39217 39224, https://doi.org/10.1074/jbc.M606100200. 27 (1999) 1087–1094, https://doi.org/10.1016/S0891-5849(99)00142-2. [4] R.H. Houtkooper, M. Turkenburg, B.T. Poll-The, D. Karall, C. Pérez-Cerdá, [28] H. Park, A. He, M. Tan, J.M. Johnson, J.M. Dean, T.A. Pietka, Y. Chen, X. Zhang, A. Morrone, S. Malvagia, R.J. Wanders, W. Kulik, F.M. Vaz, The enigmatic role of F.F. Hsu, B. Razani, K. Funai, I.J. Lodhi, Peroxisome-derived lipids regulate adipose tafazzin in cardiolipin metabolism, Biochim. Biophys. Acta Biomembr. 1788 (2009) thermogenesis by mediating cold-induced mitochondrial fission, J. Clin. Invest. 129 – 2003 2014, https://doi.org/10.1016/j.bbamem.2009.07.009. (2019) 694–711, https://doi.org/10.1172/JCI120606. [5] P. Vreken, F. Valianpour, L.G. Nijtmans, L.A. Grivell, B. Plecko, R.J.A. Wanders, [29] N. Ikon, F.F. Hsu, J. Shearer, T.M. Forte, R.O. Ryan, Evaluation of cardiolipin na- P.G. Barth, Defective remodeling of cardiolipin and phosphatidylglycerol in Barth nodisks as lipid replacement therapy for Barth syndrome, J. Biomed. Res. 32 (2018) – syndrome, Biochem. Biophys. Res. Commun. 279 (2000) 378 382, https://doi.org/ 107–112, https://doi.org/10.7555/JBR.32.20170094. 10.1006/bbrc.2000.3952. [30] M. Honsho, S. Asaoku, Y. Fujiki, Posttranslational regulation of fatty acyl-CoA re- [6] M. Schlame, J.A. Towbin, P.M. Heerdt, R. Jehle, S. DiMauro, T.J.J. Blanck, ductase 1, Far1, controls ether glycerophospholipid synthesis, J. Biol. Chem. 285 fi De ciency of tetralinoleoyl-cardiolipin in Barth syndrome, Ann. Neurol. 51 (2002) (2010) 8537–8542, https://doi.org/10.1074/jbc.M109.083311. – 634 637, https://doi.org/10.1002/ana.10176. [31] T.E. Merchant, T. Glonek, 31P NMR of tissue phospholipids: competition for Mg2+, [7] Z. Gu, F. Valianpour, S. Chen, F.M. Vaz, G.A. Hakkaart, R.J.A. Wanders, Ca2+, Na+ and K+ cations, Lipids 27 (1992) 551–559, https://doi.org/10.1007/ M.L. Greenberg, Aberrant cardiolipin metabolism in the yeast taz1 mutant: a model BF02536139. – for Barth syndrome, Mol. Microbiol. 51 (2004) 149 158, https://doi.org/10.1046/ [32] J. Folch, M. Lees, G.H. Sloane Stanley, A simple method for the isolation and j.1365-2958.2003.03802.x. purification of total lipides from animal tissues, J. Biol. Chem. 226 (1957) 497–509. [8] G. Paradies, V. Paradies, F.M. Ruggiero, G. Petrosillo, Role of cardiolipin in mi- [33] N. Li, D.Y. Lizardo, G.E. Atilla-Gokcumen, Specific triacylglycerols accumulate via tochondrial function and dynamics in health and disease: molecular and pharma- increased lipogenesis during 5-FU-induced apoptosis, ACS Chem. Biol. 11 (2016) cological aspects, Cells 8 (2019) 728, https://doi.org/10.3390/cells8070728. 2583–2587, https://doi.org/10.1021/acschembio.6b00410. [9] F. Gonzalvez, M. D'Aurelio, M. Boutant, A. Moustapha, J.P. Puech, T. Landes, L. [34] F.J. Miller, F.L. Rosenfeldt, C. Zhang, A.W. Linnane, P. Nagley, Precise determi- Arnauné-Pelloquin, G. Vial, N. Taleux, C. Slomianny, R.J. Wanders, R.H. nation of mitochondrial DNA copy number in human skeletal and cardiac muscle by Houtkooper, P. Bellenguer, I.M. Møller, E. Gottlieb, F.M. Vaz, G. Manfredi, P.X. a PCR-based assay: lack of change of copy number with age, Nucleic Acids Res. 31

8 J.C. Bozelli, et al. BBA - Molecular and Cell Biology of Lipids 1865 (2020) 158677

(2003) e61, , https://doi.org/10.1093/nar/gng060. biosynthesis and peroxisome assembly, Proc. Natl. Acad. Sci. U. S. A. 83 (1986) [35] R.L.A. Zoeller, T.J. Grazia, P. Lacamera, J. Park, D.L.P. Gaposchkin, H.W. Farber, 5170–5174, https://doi.org/10.1073/pnas.83.14.5170. Increasing plasmalogen levels protects human endothelial cells during hypoxia, Am. [43] F. Dorninger, A. Brodde, N.E. Braverman, A.B. Moser, W.W. Just, S. Forss-Petter, J. Physiol. Heart Circ. Physiol. 283 (2002) H671–H679. B. Brügger, J. Berger, Homeostasis of phospholipids - the level of phosphatidy- [36] N. Maulik, A. Tosaki, R.M. Engelman, G.A. Cordis, D.K. Das, Myocardial salvage by lethanolamine tightly adapts to changes in ethanolamine plasmalogens, Biochim. l-o-hexadecyl-sn-glycerol: possible role of peroxisomal dysfunction in ischemia re- Biophys. Acta Mol. Cell Biol. Lipids 1851 (2015) 117–128. perfusion injury, J. Cardiovasc. Pharmacol. 24 (1994) 486–492, https://doi.org/10. [44] M. Saitoh, M. Itoh, S. Takashima, M. Mizuguchi, M. Iwamori, Phosphatidyl etha- 1097/00005344-199409000-00018. nolamine with increased polyunsaturated fatty acids in compensation for plasma- [37] A.K. Das, R.D. Holmes, G.N. Wilson, A.K. Hajra, Dietary ether lipid incorporation logen defect in the Zellweger syndrome brain, Neurosci. Lett. 449 (2009) 164–167, into tissue plasmalogens of humans and rodents, Lipids. 27 (1992) 401–405, https://doi.org/10.1016/j.neulet.2008.11.004. https://doi.org/10.1007/BF02536379. [45] M. Schlame, R.I. Kelley, A. Feigenbaum, J.A. Towbin, P.M. Heerdt, T. Schieble, [38] R.D. Holmes, G.N. Wilson, A. Hajra, Oral ether lipid therapy in patients with per- R.J.A. Wanders, S. DiMauro, T.J.J. Blanck, Phospholipid abnormalities in children oxisomal disorders, J. Inherit. Metab. Dis. 10 (1987) 239–241, https://doi.org/10. with Barth syndrome, J. Am. Coll. Cardiol. 42 (2003) 1994–1999, https://doi.org/ 1007/BF01811415. 10.1016/j.jacc.2003.06.015. [39] P. Brites, A.S. Ferreira, T.F. da Silva, V.F. Sousa, A.R. Malheiro, M. Duran, [46] H. Chao, T.S. Anthonymuthu, E.M. Kenny, A.A. Amoscato, L.K. Cole, G.M. Hatch, H.R. Waterham, M. Baes, R.J.A. Wanders, Alkyl-glycerol rescues plasmalogen levels J. Ji, V.E. Kagan, H. Bayır, Disentangling oxidation/hydrolysis reactions of brain and pathology of ether-phospholipid deficient mice, PLoS One 6 (2011) e28539, , mitochondrial cardiolipins in pathogenesis of traumatic injury, JCI Insight 3 (2018) https://doi.org/10.1371/journal.pone.0028539. 1–15, https://doi.org/10.1172/jci.insight.97677. [40] S. Chung, P.P. Dzeja, R.S. Faustino, C. Perez-Terzic, A. Behfar, A. Terzic, [47] C. Ye, Z. Shen, M.L. Greenberg, Cardiolipin remodeling: a regulatory hub for Mitochondrial oxidative metabolism is required for the cardiac differentiation of modulating cardiolipin metabolism and function, J. Bioenerg. Biomembr. 48 (2016) stem cells, Nat. Clin. Pract. Cardiovasc. Med. 4 (2007) S60–S67, https://doi.org/10. 113–123, https://doi.org/10.1007/s10863-014-9591-7. 1038/ncpcardio0766. [48] M.G. Baile, K. Whited, S.M. Claypool, Deacylation on the matrix side of the mi- [41] N. Nagan, R.A. Zoeller, Plasmalogens: biosynthesis and functions, Prog. Lipid Res. tochondrial inner membrane regulates cardiolipin remodeling, Mol. Biol. Cell 24 40 (2001) 199–229, https://doi.org/10.1016/S0163-7827(01)00003-0. (2013) 2008–2020, https://doi.org/10.1091/mbc.E13-03-0121. [42] R.A. Zoeller, R.H. Raetz, Isolation of animal cell mutants deficient in plasmalogen